Announcement

Collapse
No announcement yet.

AMD AOCC 2.2 Helping Squeeze Extra Performance Out Of AMD EPYC 7002 "Rome" CPUs

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • AMD AOCC 2.2 Helping Squeeze Extra Performance Out Of AMD EPYC 7002 "Rome" CPUs

    Phoronix: AMD AOCC 2.2 Helping Squeeze Extra Performance Out Of AMD EPYC 7002 "Rome" CPUs

    At the end of June AMD quietly released a new version of the AMD Optimizing C/C++ Compiler. Noticing the new release this week, here are some benchmarks of AOCC 2.2 up against LLVM Clang 10 and GCC 10 with Ubuntu Linux while running from an AMD EPYC 7742 2P server for looking at the performance gains possible with the compiler optimizations.

    http://www.phoronix.com/vr.php?view=29464

  • #2
    “For most of the other workloads tested, AOCC 2.2 provided minor advantages on top of Clang 10.“

    But with really terrible compile time regressions (vs Clang). They should work with upstream rather than doing in house very costly passes which are not worth it as we see it here they should push useful things to llvm trunk.

    Comment


    • #3
      Originally posted by pyler View Post
      But with really terrible compile time regressions (vs Clang). They should work with upstream rather than doing in house very costly passes which are not worth it as we see it here they should push useful things to llvm trunk.
      We do work with upstream. The releases are just an easy way for customers to get access to all of the stuff we are working on.

      Comment


      • #4
        Thanks!

        Are you aware of compile time issues? Or is it not so important?

        Comment


        • #5
          GCC faster compilation times, but slower performance (much slower sometimes). Something not expected.

          Comment


          • #6
            Originally posted by pyler View Post
            “For most of the other workloads tested, AOCC 2.2 provided minor advantages on top of Clang 10.“

            But with really terrible compile time regressions (vs Clang). They should work with upstream rather than doing in house very costly passes which are not worth it as we see it here they should push useful things to llvm trunk.
            For certain applications this is perfectly tolerable. If I am compiling a binary that will be running on 120 cores 24/7 for the next three months, I will gladly take a 10x compile time regression for a couple percent boost in execution speed.

            Comment


            • #7
              This sentence is a bit confusing...

              Originally posted by phoronix View Post
              While the Optimizing "C/C++" Compiler, the Fortran language support continues being improved upon by leveraging LLVM's FLANG.

              Comment


              • #8
                Originally posted by tildearrow View Post
                This sentence is a bit confusing...
                What are you confused by?

                Comment


                • #9
                  Originally posted by schmidtbag View Post
                  What are you confused by?
                  The part in bold. Shouldn't it be "While being an Optimizing C/C++ Compiler"?

                  It sounds like:

                  While the Strawberry Juice, the Milk support continues being improved upon by leveraging a blender.
                  While SUBJECT, the SUBJECT VERB OBJECT.
                  It should be:

                  While SUBJECT VERB OBJECT, the SUBJECT VERB OBJECT
                  Last edited by tildearrow; 08-28-2020, 08:38 PM.

                  Comment


                  • #10
                    It can be further optimized as:
                    "Primary C/C++, secondary Fortran."

                    Comment

                    Working...
                    X